projects / oota-llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 7e61b1f) | patch
[Reassociate] Canonicalize operands of vector binary operators.
authorChad Rosier <mcrosier@codeaurora.org>
Fri, 14 Nov 2014 17:08:15 +0000 (17:08 +0000)
committerChad Rosier <mcrosier@codeaurora.org>
Fri, 14 Nov 2014 17:08:15 +0000 (17:08 +0000)
commit1298a29c64ac3a03c79f37e2a8ad5abd815dd93e
tree4b99dc24b6e5769bab5bf3eac3dab531e7ca5747tree | snapshot
parent7e61b1fb621d84a7e504f77e4deee206e7cd5075commit | diff
[Reassociate] Canonicalize operands of vector binary operators.

Prior to this commit fmul and fadd binary operators were being canonicalized for
both scalar and vector versions.  We now canonicalize add, mul, and, or, and xor
vector instructions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@222006 91177308-0d34-0410-b5e6-96231b3b80d8
lib/Transforms/Scalar/Reassociate.cpp diff | blob | history
test/Transforms/Reassociate/fast-ReassociateVector.ll diff | blob | history
C/C++ LLVM-based compilers that forbids OOTA behaviors